ABSTRACT There is an urgent, unmet clinical need to develop non-invasive approaches for distinguishing benign vs. malignant indeterminate pulmonary nodules (IPN) identified on CT chest. We propose to develop and validate integrated clinical, molecular and imaging-based diagnostic models of lung cancer in smokers with nodules 6- 25 mm who are at elevated risk of lung cancer as a result of meeting eligibility criteria for screening, and whose nodules may have been screen-detected or incidentally-detected in routine clinical practice. This nodule size range represents an intermediate risk for disease for which there is the greatest clinical uncertainty in terms of diagnostic management. The investigators at BU have developed and validated a gene expression biomarker, recently launched commercially as a CLIA assay (PerceptaTM) measured in cytologically-normal mainstem bronchus epithelium with high sensitivity and high negative predictive value (NPV) for detecting lung cancer among smokers undergoing bronchoscopy for suspect lung cancer. They have recently extended these cancer-specific molecular alterations within the ?field of injury? to develop and validate a similar biomarker in less invasively collected nasal epithelium. Additionally, investigators at UCLA have identified both qualitative and quantitative imaging features that inform diagnostic risk in both screen- and incidentally-detected nodules in older smokers. In Aim 1 of this proposal, we will refine qualitative and quantitative imaging biomarkers, confirm their reproducibility, and determine their contribution to diagnostic models in individuals with nodules 6- 25 mm from the CT arm of the National Lung Screening Trial (NLST). Aim 2 will determine whether bronchial gene expression biomarkers originally validated in high risk cohorts perform equally well in the specific context of patients with IPNs 6-25 mm undergoing bronchoscopy as part of the Detection of Early Lung Cancer Among Military Personnel (DECAMP) consortium, as well as integrate this biomarker with imaging-based markers from Aim 1. Given that not all IPN patients undergo bronchoscopy, Aim 2 will also validate a recently developed nasal gene-expression biomarker in this same cohort and construct models that integrate clinical, imaging, and molecular biomarkers. In Aim 3, the integrated clinical, nasal gene-expression and imaging-based biomarker will then be validated prospectively in multiple cohorts with screen- and incidentally-detected IPNs who are undergoing CT surveillance or biopsy. Our working hypothesis is that diagnostic models that integrate orthogonal feature sets of molecular biomarkers, clinical variables, and imaging features will provide the highest discrimination between benign and malignant IPNs in the 6-25 mm size range in which diagnostic uncertainty is greatest. Given the increasingly widespread implementation of lung cancer screening and dramatically increased numbers of IPNs, we anticipate that sensitive biomarkers with a high NPV would enable physicians to avoid unnecessary procedures in patients with benign disease of the lung, avoiding their associated medical risks and economic costs.